paradite

MCP is basically commoditizing SaaS and software by abstracting them away behind the AI agent interface.

It benefits MCP clients (ChatGPT, Claude, Cursor, Goose) more than the MCP servers and the service behind the MCP servers (GitHub, Figma, Slack).

I see what you mean. It is a paradigm shift indeed if you look from the user's perspective.

I think this is a good explanation on the client side of MCP. But most developers are not building MCP clients (I think?). Only a few companies like OpenAI, Anthropic, Cursor and Goose are building MCP client.

Most developers are currently building MCP servers that wrap a 3rd party or wrap their own service. And in this case, they are still at deciding on the tools in design-time, not runtime.

Also I want to mention that both Cursor and Claude desktop don't support dynamic toggling on / off tools within a MCP server, which means users can't really pick which tools to expose to AI. It exposes all tools within a MCP server in current implementation.

I actually tried it a few months ago and I still using it (mainly for testing its capabilities), despite it not performing up to my standards.

I wrote my first impressions of Devin here: https://thegroundtruth.substack.com/p/devin-first-impression...

Missing a few. Check out mine visualized in 2D quardants:

https://paradite.github.io/ai-coding/

Also there are a lot of cli tools in this space:

https://prompt.16x.engineer/cli-tools

Yes. This is my point I'm trying to make. Thank you for explaining it.

Ok looks like people are not getting my comment.

Being a judge in a hackathon is one of the criterion for O-1 visa.

https://www.linkedin.com/pulse/getting-o-1-visa-easier-than-...

O-1

You are asking the right question, but to the wrong person.

Normalizing mediocrity is not something I want in STEM fields.

We need exceptional people who can push boundaries and do exceptional work in order to progress.

I don't know. As someone not from US, it looks like Sam Altman is also in good relationship with Trump due to Stargate.

My burning question: Why not also make a slightly larger model (100B) that could perform even better?

Is there some bottleneck there that prevents RL from scaling up performance to larger non-MoE model?

Interesting. This is the first time I am hearing about intrinsic positional bias for LLM. I had some intuition on this but nothing concrete.

Thinking from the retrieval perspective, would it make sense to have two layers?

First layer just describes on high level, the tools available and what they do, and make the model pick or route the request (via system prompt, or small model).

Second layer implements the actual function calling or OpenAPI, which then would give the model more details on the params and structures of the request.

Hi. This is very helpful. Thanks for sharing!

Hi. I'm an electron app developer. I use electron builder paired with AWS S3 for auto update.

I have always put Windows signing on hold due to the cost of commercial certificate.

Is the Azure Trusted Signing significantly cheaper than obtaining a commercial certificate? Can I run it on my CI as part of my build pipeline?

It's actually fairly easy to setup a 3rd party app to use Claude via API, to get extremely generous limits.

I wrote a step-by-step guide for the app I built: https://prompt.16x.engineer/guide/claude

It's there in the docs (Model comparison table) https://docs.anthropic.com/en/docs/about-claude/models/all-m...

DeepSeek also has a FIM (Fill In the Middle) completion model via API, if anyone is interested to try out:

https://api-docs.deepseek.com/guides/fim_completion

Can you save the full request details and full response details in one click as one file with this extension?

This is the feature I use the most.

Thanks a lot for the explanation.

If I can ask one more question, why don't Ollama use binaries of pre-built llama.cpp with Vulkan support directly?

At the risk of being pedantic (I don't know much about C++ and I'm genuinely curious), if Ollama is really just a wrapper around Llama.cpp, why would it need the Vulkan specific flags?

Shouldn't it just call Llama.cpp and let Llama.cpp handle the flags internally within Llama.cpp? I'm thinking from an abstraction layer perspective.

I guess git and GitHub are working as intended then.

This is not a sarcastic comment. I'm genuinely happy that this was the outcome.

Ok I was looking at the repo from mobile and missed the releases.

Still it's not immediate obvious from README that there is an option to download it. There are instructions on how to build it, but not how to download it. Or maybe I'm blind, please correct me.

Ok assuming what you said is correct, why wouldn't Ollama then be able to support Vulkan by default out of the box?

Sorry I'm not sure what's the relationship exactly between the two projects. This is a genuine questions, not a troll question.

Could it be that supporting multiple platforms open up more support tickets and adds more work to keep the software working on those new platforms?

As someone who built apps for Windows, Linux, macOS, iOS and Android, it is not trivial to ensure your new features or updates work on all platforms, and you have to deal with deprecations.

For starters:

- It doesn't have a website

- It doesn't have a download page, you have to build it yourself

It's worth mentioning the fact that you need to adjust the context window manually for coding tasks (default 2k is not enough for coding tasks).

Here's how to run deepseek-r1:14b (DeepSeek-R1-Distill-Qwen-14B) and set it to 8k context window:

  ollama run deepseek-r1:14b
  /set parameter num_ctx 8192
  /save deepseek-r1:14b-8k

  ollama serve

Haha no worries. This is a perfectly valid use case of LLM. I'm happy that the comment sounds very professional and to the point.

Btw completely off topic, but your comment triggered the internal classification in my brain, and it looks like AI-generated.

Not accusing you anything. Could be that you happen to write in a way similar to LLMs. Could be that we are influenced by LLM writing styles and are writing more and more like LLMs. Could be that the difference between LLM generated content and human-generated content is getting smaller and harder to tell.